ABSTRACT

Statistical methods (SM) have been dominant in generating insights from any type of data for generations. However, with the recent advances in technology, machine learning (ML) has become one of the widely spoken methods to generate insights with more ease of use. While the followers of statistical methods have a differing view point about ML, and the followers of ML have a differing viewpoint about SM, this article isolates the merits of each of these two methods and advances arguments for when to use what method based on the purpose, context, frequency of use, cost, expertise and time. To be specific, the main purpose of SM is for inference and that of ML is prediction. Further, this article goes one step further and creates a scenario where it shows that when we combine the learning from using a statistical method and apply it to machine learning, the ultimate benefit can be greater than the sum of each method’s benefits.

摘要

统计方法 (SM) 在从几代人的任何类型的数据中产生洞察力方面一直占主导地位。然而，随着最近技术的进步，机器学习 (ML) 已成为广泛使用的方法之一，可以更轻松地生成洞察力。虽然统计方法的追随者对 ML 有不同的看法，而 ML 的追随者对 SM 有不同的看法，但本文将这两种方法的优点分开，并根据目的提出何时使用哪种方法的论点、背景、使用频率、成本、专业知识和时间。具体来说，SM的主要目的是推理，ML的主要目的是预测。更多，